Water intoxication is a serious problem in many patients with chronic psychiatric illness. In an effort to determine the mechanism of this disorder, we investigated the osmoregulation of water intake and antidiuretic function in psychiatric patients with polydipsia and hyponatremia and in matched controls with psychiatric illness but neither polydipsia nor hyponatremia. We found that a water load suppressed plasma osmolality and vasopressin and urine osmolality in both groups, but that urinary dilution and free water clearance were impaired in the patients with hyponatremia, even though plasma levels of vasopressin and solute clearance were similar in the two groups. Moreover, during water loading and infusion of hypertonic saline, the plasma level of vasopressin was higher at any given plasma osmolality in the test patients than in the controls, indicating a downward resetting of the osmostat. Patients' estimates of the amount of water they desired were shown to correlate significantly with the amount of water consumed and, at any given level of plasma osmolality, appeared to be higher in the test patients than in the controls. We conclude that psychiatric patients with polydipsia and hyponatremia have unexplained defects in urinary dilution, the osmoregulation of water intake, and the secretion of vasopressin.

This study describes the prevalence of violence and the putative risk factors for violence in 184 Alzheimer patients and their primary caregivers living in the community. Analysis of the severe violence subscale of the Conflict Tactics Scale indicated that 15.8% of patients had been violent in the year since diagnosis. A total of 5.4% of caregivers reported being violent toward the patient. The overall prevalence of violence was 17.4%. The variables most associated with violence were caregiver depression and living arrangement.

Our hospice enrollment criteria identified a group with a median survival time of 4 months and a mean survival time of 6.9 months. Using NHO criteria relying on the FAST allows the identification of a subgroup with very high mortality and a short time until death. Although the FAST can identify a subgroup of appropriate candidates for hospice, sole reliance on this measure might decrease access to hospice care for many dementia patients.

A national survey of hospices revealed that few patients with primary dementia are currently treated by these programs, unless they have other terminal illnesses. An inability to predict survival was the major reason offered to explain this phenomenon. The higher percentage of patients with secondary dementia in hospice suggests that dementia per se does not make hospices care unfeasible. Similarly, the high proportion of for-profit hospices that enrolled patients whose dementia was primary implies the fiscal feasibility of providing hospice care for these patients. Further study is needed concerning the determinants of survival time in end-stage dementia.